Interaction between Sterol Regulatory Element-binding Proteins and Liver Receptor Homolog-1 Reciprocally Suppresses Their Transcriptional Activities*

In previous studies it was demonstrated that sterol regulatory element-binding proteins (SREBPs) are able to interact with one of the nuclear receptors, hepatocyte nuclear receptor (HNF)-4, and that this interaction regulates transcriptional activities of these proteins (Misawa, K., Horiba, T., Arimura, N., Hirano, Y., Inoue, J., Emoto, N., Shimano, H., Shimizu, M., and Sato, R. (2003) J. Biol. Chem. 278, 36176–36182; Yamamoto, T., Shimano, H., Nakagawa, Y., Ide, T., Yahagi, N., Matsuzaka, T., Nakakuki, M., Takahashi, A., Suzuki, H., Sone, H., Toyoshima, H., Sato, R., and Yamada, N. (2004) J. Biol. Chem. 279, 12027–12035). In an attempt to identify other nuclear receptor family members affecting the SREBP transcriptional activities, we found that the liver receptor homolog (LRH)-1 suppresses them. Several types of luciferase assays revealed that coexpression of these two proteins (LRH-1 and SREBP-1a, -1c, or -2) results in reciprocal inhibition of the transcriptional activity of each protein. It was confirmed that suppression in endogenous LRH-1 by small interference RNA stimulates the mRNA levels of certain SREBP target genes and that elevation in active SREBPs in the nucleus in response to cholesterol depletion suppresses the LRH-1 activity. In vitro/in vivo glutathione S-transferase pulldown experiments demonstrated that the basic helix-loop-helix-leucine zipper domain in SREBP-2 binds to the ligand-binding domain in LRH-1. Furthermore, we found that SREBP-2 interferes with the recruitment of a coactivator of LRH-1, the peroxisome proliferator-activated receptor γ coactivator-1α, thereby leading to the inhibition of the LRH-1 transcriptional activity. These results clearly indicate that the interaction between SREBPs and LRH-1 exerts a suppressive influence on their target gene expression responsible for cholesterol and bile acid metabolism.

SREBPs 3 are synthesized as membrane-bound precursor proteins and proteolytically processed to yield the N-terminal transcription factor domain that enters the nucleus. In the nucleus these transcription factors activate most genes required to produce cholesterol and fatty acids. As long as intracellular cholesterol concentrations are sufficient, SREBPs remain bound to mainly the endoplasmic reticulum (ER) as a trimer composed of SREBP, the SREBP cleavage-activating protein, and the insulin-inducing gene. Once the cholesterol contents in ER membranes decline, the SREBP⅐SREBP cleavage-activating protein complex can no longer bind to insulininducing gene, thereby exiting the ER and reaching the Golgi apparatus where the SREBPs are proteolytically processed by two Golgi-associated membrane bound proteases (1,2).
Concomitant with the well regulated proteolytic activation of SREBPs on the ER-Golgi membranes, the transcriptional activities of nuclear SREBPs are also modulated in diverse ways. We previously reported that the nuclear forms of SREBPs are rapidly degraded after being modified by polyubiquitin chains through a ubiquitin-proteasome pathway (3). The strict activation of SREBPs by the proteolytic process through monitoring the ER membrane cholesterol content becomes relevant only when the nuclear SREBPs are not retained in the nucleus for a long period. Sumoylation or phosphorylation is another type of modification of nuclear SREBPs conditioning their transcriptional activities (4 -7). Moreover, the interaction with other nuclear proteins such as the cAMP response element-binding protein (CREB) and HNF-4 can also modulate SREBP activities (8,9). Although the interaction between SREBPs and certain transcription factors is thought to exert reciprocal effects on their transcriptional activities, such cross-talk remains poorly understood.
In the present study we demonstrate that LRH-1, among several nuclear receptors involved in the regulation of lipid metabolism, affects the SREBP family members' transcriptional activities. This cross-talk brings about the mutual inhibition of transcriptional activity. We demonstrate the protein interaction between LRH-1 and SREBPs and identify the interaction regions of these molecules. Based on the fact that the activities of several nuclear receptors are modulated by the small het-erodimer partner (SHP), an LRH-1 target gene (10 -15), the novel regulatory system mediated by the interaction between LRH-1 and SREBPs very likely plays a pivotal role in maintaining lipid homeostasis.
Reporter Assays-Reporter assays were performed as described previously (18,19). HEK293 cells (35-mm dishes) were transfected by the calcium phosphate method with 0.2 g of a reporter plasmid, 0.01 g of phRL-TK, an expression plasmid encoding Renilla luciferase (Promega), and 0.1 g of an expression plasmid. HepG2 cells (35-mm dishes) were trans-fected by the calcium phosphate method with 2 g of a reporter plasmid, 0.1 g of phRL-TK, and 2 or 6 g of the SREBP-2 expression plasmid. Huh7 cells (20-mm dishes) were transfected by using Lipofectamine Transfection Reagent (Invitrogen) with 0.2 g of a reporter plasmid, 0.1 g of pGal4-LRH1, 0.01 g of phRL-TK, and various amounts (0.1 and 0.3 g) of an SREBP expression plasmid. Forty-eight hours later, both firefly and Renilla luciferase activities were quantified using a Dual-Luciferase TM reporter system (Promega) according to the manufacturer's instructions.
In Vitro GST Pulldown Assays-The fusion proteins, GST-SREBP2-(1-481) and GST-LRH-1, and GST were expressed in Escherichia coli and bound to glutathione-Sepharose beads 4B (GE Healthcare Life Sciences) following the manufacturer's instructions. Using the TNT coupled transcription/translation kit (Promega), 35 S-labeled LRH-1 or SREBP-2 was synthesized according to the manufacturer's protocol. Twenty-five microliters of either GST or the GST fusion protein-coupled resin, 10 l of the in vitro translation reaction, and 300 l of binding buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM EDTA, 0.5% Nonidet P-40, 1 l/ml protease inhibitor mixture (Sigma)) were incubated at 4°C for 30 min. The reactions were washed five times with 500 l of buffer, and the specifically bound proteins were pelleted, resuspended in sample buffer, and analyzed by SDS-PAGE followed by autoradiography. The signals on the membrane were detected using an image-analyzing system (FLA-3000, Fujifilm Inc.).
In Vivo GST Pulldown and Immunoprecipitation Experiments-HepG2 cells (100-mm dishes) were transfected with the indicated plasmids (5 g each) by the calcium phosphate method. Forty-eight hours later, the cells were harvested, and whole cell lysates were subjected to either GST pulldown assays or immunoprecipitation using anti-FLAG antibodies (M2, Sigma). Western blot analysis was carried out using anti-GST (Sigma), anti-FLAG, anti-Myc (Sigma), or anti-SREBP2 antibodies RS004 (16) with chemiluminescent substrate (ECL, GE Healthcare Life Sciences). The signals were detected with a LuminoImager (LAS-3000, Fujifilm).
Real-time PCR-Huh7 cells (20-mm dishes) were transfected with siRNA oligonucleotides (10 pmol) using Lipofectamine 2000 Reagent. Seventy-two hours later, total cellular RNA was extracted and reverse transcribed with Superscript III (Invitrogen). Fluorescence real-time PCR was performed using Taq-Man Gene Expression Assays (Applied Biosystems) on an ABI PRISM 7000 system. S17 rRNA protein transcript was used as an internal control to normalize the variations of the RNA amounts.

LRH-1 and SREBPs Inhibit the Transcriptional Activities of
SREBPs and LRH-1, Respectively-In the previous study we employed a heterologous Gal-4 system using expression plasmids encoding an active form of SREBP-1 or -2 coupled to the DNA-binding domain (DBD) of yeast Gal4, to determine the direct effect of HNF-4 on the transcriptional activity of SREBPs (9). Using this assay system we examined whether other nuclear receptors exert any effects through a mutual interaction and found that LRH-1, unlike HNF-4, significantly suppresses the transcriptional activity of SREBP-1 and -2 in a dose-dependent manner (Fig. 1, A and B). To further confirm the function of LRH-1, the effect of LRH-1 on the SREBP-induced promoter activity of SREBP-responsive genes was analyzed using the native promoter reporter genes. Fig. 1 (C and D) indicates that the promoter activities of both HMG-CoA synthase and LDL receptor induced by either SREBP-1a or SREBP-2 were dosedependently suppressed in the presence of LRH-1. The possibility that LRH-1 directly acts on the specific motif in these native promoters can be ruled out, because the promoter activities were not affected by LRH-1 alone (Fig. 1C). Judging from the finding that the addition of LRH-1 almost completely reduced the luciferase activities induced by Gal4-SREBPs (Fig.  1, A and B), it is likely that LRH-1 directly inactivates the transactivation property of SREBPs. Therefore, the simplest explanation is that the inhibitory effect of LRH-1 observed in Fig. 1 (C and D) was not due to a reduction in the SREBP binding to SREBP-responsive elements located in these promoters.
To examine whether SREBPs affect the LRH-1 transcriptional activity, we generated another construct encoding LRH-1 coupled to the DNA-binding domain of yeast Gal4. Compared with Gal4-SREBPs, the transactivation property of this fusion protein was relatively low in several cell lines, and we finally found that human hepatoma Huh7 cells gave us the best response. In this experiment we examined whether SREBP-1c, an alternative splicing isoform of SREBP-1 predominantly observed in various tissues, has a similar activity to SREBP-1a, which is a more potent transcription factor with a longer N-terminal transactivation domain. As shown in Fig. 1E, in the presence of any of the SREBP family members the transcriptional activity of LRH-1 was largely inhibited. These results indicate that SREBP family members almost equally suppress the transactivation property of LRH-1. Based on this finding and the fact that SREBP-1a, rather than SREBP-1c, is the predominant isoform expressed in most culture cell lines, we used mainly SREBP-1a and SREBP-2 in the following experiments.
Next, we examined the inhibitory effect of SREBPs on the LRH-1-induced promoter activity of the human SHP gene, which contains some LRH-1 response elements as well as SREBP-responsive elements (22). When HEK293 cells lacking endogenous LRH-1 were transfected with the human SHP promoter gene together with an expression plasmid of LRH-1, a full induction of luciferase activities was observed (Fig. 1F). Consistent with a previous report (22), expression of SREBP1a-(1-487), but not SREBP2-(1-481), also stimulated the promoter activity, although the induction was more modest than that driven by LRH-1. Both SREBP-1a and SREBP-2 inhibited the LRH-1-mediated induction of the SHP promoter activity in a dose-dependent manner. These results clearly demonstrate that LRH-1 is the predominant transcription factor regulating human SHP promoter activity and that SREBPs act as negative regulators of SHP gene expression.
LRH-1 Gene Knockdown Affects the SREBP Activities-We next suppressed endogenous LRH-1 gene expression in Huh7 cells using an LRH-1-specific siRNA and examined the SREBP inhibitory effects on the SHP promoter activity. In the presence of a control siRNA the SHP promoter activity declined as the result of overexpression of the active form of SREBP-2 in a dose-dependent manner ( Fig. 2A), consistent with the results observed in HEK293 cells overexpressing LRH-1 (Fig. 1F). When the endogenous LRH-1 gene expression was suppressed with a gene-specific siRNA, SHP promoter activity was reduced even in the absence of SREBP-2, and remained at the repressed steady-state level independent of the presence of SREBP-2. These results suggest that SREBP-2 inhibited the endogenous LRH-1 activity that induces SHP gene expression.
Moreover, we examined the effect of reduced LRH-1 gene expression by the siRNA on the mRNA levels of the SREBPresponsive genes. Human hepatoma Huh7 cells cultured under cholesterol-depleted conditions to activate SREBPs were treated with siRNAs. Reduced LRH-1 gene expression by the siRNA directed against LRH-1 resulted in an increase in mRNA levels of the LDL receptor and fatty acid synthase genes, which had been stimulated by activated SREBPs under the cholester- A and B, HEK293 cells were transfected with either 0.1 g of pGAL4-SREBP1a (Gal4-DBD-SREBP1a) or -SREBP2 (Gal4-DBD-SREBP2), 0.2 g of pG5Luc containing five copies of the Gal4 binding sites, and 10 ng of phRL-TK together with increasing amounts of an expression vector for HNF-4␣ or LRH-1 (0.2 and 0.6 g); they were then cultured with a medium containing 10% FBS for 48 h. Luciferase assays were performed as described under "Experimental Procedures." The promoter activities in the absence of pGAL4-SREBP1a or -SREBP2 are represented as 1. All data are presented as means Ϯ S.D. and represent at least three independent experiments performed in triplicate. C and D, HEK293 cells were transfected with either 0.2 g of pHMG-syn containing the human HMG-CoA synthase promoter or pLDLR containing the human LDL receptor promoter, either 0.1 g of pSREBP1a-(1-487) or pSREBP2-(1-481), 10 ng of phRL-TK, and increasing amounts (0.2 and 0.6 g) of pTarget-LRH1, an expression plasmid for human LRH-1; they were then cultured with a medium containing 10% FBS for 48 h. Luciferase assays were performed as described under "Experimental Procedures." The promoter activities in the absence of SREBPs and LRH-1 are represented as 1. All data are presented as means Ϯ S.D. and represent at least three independent experiments performed in triplicate. E, Huh7 cells were transfected with pGAL4-LRH-1 (0.1 g), pG5Luc (0.2 g), and phRL-TK (10 ng) together with increasing amounts (0.1 and 0.3 g) of one of the expression vectors for SREBP-1a-(1-487), -1c-(1-457), and -2-(1-481); they were then cultured with a medium containing 10% FBS for 48 h. Luciferase assays were performed as described under "Experimental Procedures." The promoter activities in the absence of SREBPs are represented as 100%. All data are presented as means Ϯ S.D. and represent at least three independent experiments performed in triplicate. F, HEK293 cells were transfected with 0.2 g of pSHP-1300 (a reporter gene containing the human SHP promoter), 0.1 g of pTarget-LRH-1, 10 ng of phRL-TK and increasing amounts (0.2 and 0.6 g) of either pSREBP1a-(1-487) or pSREBP2-(1-481); they were then cultured with a medium containing 10% FBS for 48 h. Luciferase assays were performed as described under "Experimental Procedures." The promoter activities in the absence of SREBPs and LRH-1 are represented as 1. All data are presented as means Ϯ S.D. and represent at least three independent experiments performed in triplicate. The inset shows the sites of several LRH-1 response elements (LRHREs) and sterol regulatory elements in the human SHP promoter (22). ol-depleted conditions, and a decrease in those of SHP and CYP7A1, LRH-1 target genes (Fig. 2B). Importantly, endogenous SREBPs functionally associate with endogenous LRH-1, leading to a reciprocal inhibition.
Activity of Endogenous LRH-1 in Huh7 Cells Is Attenuated by SREBPs-To examine the effect of SREBPs on endogenous LRH-1 activities, we carried out the following experiments. Huh7 cells expressing endogenous LRH-1 were transfected with a reporter plasmid containing the native SHP promoter and cultured under either cholesterol-depleted or -loaded conditions to alter the active SREBP levels in the nucleus. When cells were cultured under the cholesterol-depleted conditions, which brought about an increase in the nuclear active SREBPs and thereby stimulated HMG-CoA synthase promoter activity, the SHP promoter activity induced by endogenous LRH-1 was suppressed (Fig. 3A). Moreover, the mRNA levels of endogenous LRH-1 target genes, SHP, CYP7A1, and apolipoprotein A-I (23), were also reduced when Huh7 cells were cultured under the cholesterol-depleted conditions, whereas the LDL receptor mRNA level was largely raised (Fig. 3B). These results indicate that endogenous LRH-1 and SREBPs vigorously interact in modulating the expression of their target genes in hepatocytes.
Basic Helix-Loop-Helix Leucine Zipper Domain of SREBPs Can Bind to the DNA-binding Domain of LRH-1-The above data strongly suggest a mutual interaction between the two proteins to fulfill their functions. To gain direct evidence of such protein-protein interaction, GST pulldown assays were performed. Several recombinant SREBP-2 isoforms lacking the N-terminal region were generated and analyzed in GST pulldown assays. Because it has been elucidated that the acidic N-terminal region of SREBPs acts as the transcriptional activa-tion domain in recruiting a coactivator, the CREB-binding protein (CBP) (24), it is conceivable that LRH-1 might interact with this portion of SREBPs, competing with CBP. Contrary to our expectation, however, several truncated versions of SREBP-2 (#1 to #5) as well as wild-type SREBP-2 bound to a GST-LRH-1 fusion protein (Fig. 4, A and B). 80 amino acid residues in the C terminus (SREBP2-(1-401), Fig. 4, A and C, #6) were not required, but 70 amino acid residues in the bHLH-Zip domain (Fig. 4, A and C, #7) were indispensable. Because wild-type SREBP-1a also binds to LRH-1 (Fig. 4, A and C, #8), it is likely that the bHLH-Zip domain of SREBP-1a and SREBP-2, which have 71% amino acid residue homology, is involved in the association with LRH-1. Similarly, the N-or C-terminal portion of LRH-1 was deleted to determine the essential intramolecular portion for the binding. GST pulldown assays revealed that deletion of the DNA-binding domain abolished the interaction with SREBP-2 (Fig. 4, D and E, #3 and #4). Taken together, it is concluded that the bHLH-Zip domain of SREBPs can be directly bound to the DNA-binding domain of LRH-1.
To assess the interaction between LRH-1 and SREBP-2 on the promoter DNA, we next performed ChIP assays using an anti-FLAG antibody. Because this antibody provided us with the most reliable and repeatable results in the current ChIP assays, we used two types of FLAG-tagged proteins, FLAG-SREBP2 and FLAG-LRH-1. When HEK293 cells were trans-FIGURE 2. LRH-1 gene knockdown affects the SREBPs. A, Huh7 cells were transfected with pSHP-1300 (30 ng), either 3 pmol of a control siRNA (siCon) or siRNA targeting LRH-1 (siLRH-1), phRL-TK (3 ng), and increasing amounts (30 and 90 ng) of pSREBP2-(1-481); they were then cultured with a medium containing 10% FBS for 48 h. Luciferase assays were performed as described under "Experimental Procedures." The promoter activities in the absence of SREBP-2 and the presence of a control siRNA are represented as 100%. All data are presented as means Ϯ S.D. and represent at least three independent experiments performed in triplicate. B, Huh7 cells were transfected with either 10 pmol of a control siRNA (siCon) or siRNA targeting LRH-1 (siLRH-1); they were then cultured with a medium containing 5% lipoprotein-deficient serum supplemented with 50 M of a HMG-CoA reductase inhibitor, pravastatin, plus 50 M sodium mevalonate for 96 h to increase in the amount of active nuclear SREBPs and total RNA was isolated. Real-time PCR analysis was performed, and relative mRNA levels were obtained after normalizing to S17 mRNA. The mRNA levels (LRH-1, LDL receptor (LDLR), fatty acid synthase (FAS), SHP, and CTP7A1) in the presence of a control siRNA (siCon) are represented as 1. All data are presented as means Ϯ S.D. and represent at least three independent experiments performed in triplicate. *, p Ͻ 0.05. FIGURE 3. SREBPs diminish the SHP promoter activity driven by endogenous LRH-1. A, Huh7 cells were transfected with either pSHP-1300 or pHMGsyn containing the human HMG-CoA synthase promoter (0.2 g) and phRL-TK (10 ng); they were then cultured with medium containing 5% lipoprotein-deficient serum supplemented with either 1 g/ml 25-hydroxycholesterol plus 10 g/ml cholesterol (cholesterol ϩ) or 50 M of a HMG-CoA reductase inhibitor, pravastatin, plus 50 M sodium mevalonate (cholesterol Ϫ) for 48 h to alter the amount of active SREBPs in the nucleus. Luciferase assays were performed as described under "Experimental Procedures." The promoter activities under sterol-loaded conditions are represented as 100%. All data are presented as means Ϯ S.D. and represent at least three independent experiments performed in triplicate. *, p Ͻ 0.05; **, p Ͻ 0.01. B, Huh7 cells were cultured under cholesterol-loaded (cholesterol ϩ) or -depleted (cholesterol Ϫ) condition for 48 h as described above, and then total RNA was isolated. Real-time PCR analysis was performed and relative mRNA levels were obtained after normalizing to S17 mRNA. The mRNA levels (SHP, CYP7A1, apolipoprotein A-I (Apo A-I), and LDL receptor (LDLR)) under cholesterolloaded condition are represented as 1. All data are presented as means Ϯ S.D. and represent at least two independent experiments performed in triplicate. *, p Ͻ 0.05; **, p Ͻ 0.01.
fected with an expression plasmid for FLAG-LRH-1, PCR products for the HMG-CoA synthase promoter were not detected (Fig. 5B), whereas the CYP7A1 promoter PCR products were amplified whenever FLAG-LRH-1 was expressed. In the pres-ence of SREBP2-(1-481), the HMG-CoA synthase promoter PCR products were found, indicating that an SREBP-2⅐LRH-1 complex bound to this promoter as illustrated in Fig. 5B (the left  upper panel). When cells expressed FLAG-SREBP2, the HMG-CoA promoter PCR products were found. Only when both FLAG-SREBP2 and LRH-1 were expressed the CYP7A1 PCR products were amplified, suggesting that an LRH-1⅐SREBP2 complex bound to the promoter as illustrated in the right bottom panel in Fig. 5B. Taken together, LRH-1 and SREBP2-(1-481) can form a functional complex on these promoters.

N-terminal Region of SREBP-2 Is Necessary for Suppression of the LRH-1 Transcriptional Activity Potentiated by the Coacti-
vator PGC-1␣-Based on our initial expectation that LRH-1 would directly interact with the N-terminal region of SREBPs, the transcriptional activation domain, thereby inhibiting the SREBPs activities, we performed reporter assays in conjunction with in vitro GST pulldown assays (Fig. 4) to verify the necessity of this terminal portion. Although wild-type SREBP2-(1-481) suppressed the SHP promoter activity driven by LRH-1 dosedependently, truncation of the N-terminal portion (SREBP2-(31-481) and SREBP2-(241-481)) completely abolished this   330)), and harvested after 48 h. One-fifth of total cell lysates were analyzed by SDS-PAGE followed by Western blotting (WB) using anti-FLAG antibodies. GST proteins attached to glutathione-Sepharose 4B were analyzed by SDS-PAGE followed by Western blotting using either anti-GST or anti-FLAG antibodies. B, ChIP assay. HEK293 cells were transfected with expression plasmids for SREBP2 and FLAG-LRH-1 or for LRH-1 and FLAG-SREBP2 and processed for ChIP analyses as described under "Experimental Procedures." After the immunoprecipitation (IP) with an anti-FLAG antibody, PCR was performed with the primers for the human HMG-CoA synthase promoter containing two sterol regulatory elements (SREs) or the human CYP7A1 promoter containing an LRHRE. 1% of the samples without the immunoprecipitation was subject to PCR as a control (1% input). effect, as initially expected (Fig. 6A). Similarly, another truncated form of SREBP-2 (SREBP2-(1-330)) did not affect the SHP promoter activity because of no interaction with LRH-1 as shown in Fig. 4C. Using the LDL receptor promoter we confirmed that both SREBP2-(31-481) and SREBP2-(241-481) worked as dominant negative forms (Fig. 6B), consistent with our previous finding (16). These results present evidence that the N-terminal region is indispensable for the inhibitory effect of SREBP-2 on the LRH-1 transcriptional activity. However, the results obtained by in vitro GST pulldown assays (Fig. 4) clearly indicate that the N-terminal region is not directly involved in the interaction. We therefore speculated that the binding of intact SREBPs to LRH-1 could interfere with the recruitment of coactivators such as PGC-1␣, or the steroid receptor coactivator-1, that potentiate the transactivation of LRH-1 (25,26). Reporter assays using the native SHP promoter revealed that PGC-1␣ augmented the LRH-1 transcriptional activity (Fig. 6C). In the presence of SREBP2-(1-481) PGC-1␣ only slightly stimulated luciferase activity, whereas the potentiation of LRH-1 activity by this coactivator was not interfered by SREBP2- . To further confirm the interference by the N-terminal portion of SREBP-2, we carried out another reporter assay using an artificial promoter containing three repeats of LRH-1 response elements. PGC-1␣ overexpression tremendously increased the promoter activity driven by LRH-1 (Fig. 6D). Wild-type SREBP-2 almost completely abolished this PGC-1␣ effect, although SREBP2-(31-481) had no effects. Moreover, in vivo immunoprecipitation experiments were performed with HepG2 cells expressing Myc-LRH-1 and FLAG-PGC-1␣. The immunoprecipitation assays revealed that interaction between PGC-1␣ and LRH-1 was noticeably inhibited by SREBP2-(1-481), but not by SREBP2-(31-481) (Fig. 6E, lanes 3-5). These results demonstrate that the N-terminal region of SREBP-2, which is not essential for the interaction with LRH-1, plays an important role in regulating the LRH-1 transcriptional activity through competing with PGC-1␣. All data are presented as means Ϯ S.D. and represent at least three independent experiments performed in triplicate. B, HEK293 cells were transfected with 0.2 g of pLDLR, 0.1 g of an expression plasmid for SREBP-2, and 10 ng of phRL-TK; they were then cultured with a medium containing 10% FBS for 48 h. Luciferase assays were performed as described under "Experimental Procedures." The promoter activities in the absence of SREBPs are represented as 1. All data are presented as means Ϯ S.D. and represent at least three independent experiments performed in triplicate. C, HEK293 cells were transfected with 0.2 g of pSHP-1300, 0.3 g of pFLAG-LRH-1, 0.1 g of pCMXPGC-1␣ (an expression plasmid for mouse PGC-1␣), 10 ng of phRL-TK, and increasing amounts (0.02 and 0.1 g) of either pSREBP2-(1-481) or -(31-481); they were then cultured with a medium containing 10% FBS for 48 h. Luciferase assays were performed as described under "Experimental Procedures." The promoter activities in the absence of SREBPs, PGC-1␣, and LRH-1 are represented as 1. All data are presented as means Ϯ S.D. and represent at least three independent experiments performed in triplicate. D, HEK293 cells were transfected with 0.2 g of pLRHREx3 (an artificial reporter gene containing three repeats of functional LRH-1 response elements), 0.1 g of pFLAG-LRH-1, 0.2 g of pCMXPGC-1␣, 10 ng of phRL-TK, and increasing amounts (0.1 and 0.3 g) of either pSREBP2-(1-481) or -(31-481); they were then cultured with a medium containing 10% FBS for 48 h. Luciferase assays were performed as described under "Experimental Procedures." The promoter activities in the absence of SREBP-2, PGC-1␣, and LRH-1 are represented as 1. All data are presented as means Ϯ S.D. and represent at least three independent experiments performed in triplicate. E, immunoprecipitation (IP) of a PGC-1␣⅐LRH-1 complex in HepG2 cells. HepG2 cells were transfected with expression plasmids (5 g each of pCMX-PGC-1␣, pMyc-LRH1, and pSREBP2-(1-481) or -(31-481)), and then harvested after 48 h. One-fifth of total cell lysates were analyzed by SDS-PAGE followed by Western blotting (WB) using either anti-Myc or anti-SREBP2 antibodies. Immunoprecipitates with anti-FLAG antibodies attached to protein A-Sepharose CL-4B were analyzed by SDS-PAGE followed by Western blotting using either anti-Myc or anti-FLAG antibodies.

DISCUSSION
Recent studies have demonstrated that the mammalian nuclear receptors SF-1 and LRH-1, sharing 56% identity in their ligand-binding domains, interact with phosphatidylcholine, phosphatidylethanolamine, phosphatidylglycerol, and phosphatidylinositols (27,28). One can speculate that the transcriptional activity of LRH-1 is differently modulated by distinct phospholipids in response to physiological changes in cell membranes. However, it seems more likely that phospholipids are used as readily available ligands, not as hormone-like signals, and that LRH-1 is constitutively active under normal physiological conditions. A known regulation system controlling the LRH-1 activity is turned on by the inhibitory action of another nuclear receptor, SHP, which is a target gene of the farnesoid X receptor that is activated by bile acids. SHP is one of two unusual orphan receptors that lack a conventional DNA-binding domain and has been reported to act as a corepressor of several nuclear receptors, including the liver X receptor and HNF-4 (10 -15). It has been demonstrated that SHP interacts with the ligand-binding domain of certain nuclear receptors through two LXXLL-related motifs, present in the N-terminal and central parts of SHP, thereby competing with coactivators that stimulate the transcriptional activity of nuclear receptors (29). It has also been demonstrated that a prospero-related homeobox transcription factor, Prox-1, with two LXXLL motifs, also suppresses the transcriptional activity of LRH-1 through interaction with it (30). Here we show that SREBPs are novel inhibitory nuclear factors modulating LRH-1 transcriptional activity without the LXXLL motif.
Although the current study revealed that LRH-1 suppresses the SREBP transcriptional activities and vice versa, we speculate that the inhibitory effects of SREBPs on the LRH-1 activity would be even more relevant than the LRH-1 inhibitory effect on the SREBP functions, because relative abundance of active SREBPs in the nucleus more dramatically changes in response to various metabolic alterations. It has been reported that a variety of factors such as insulin, cholesterol, and polyunsaturated fatty acids affect SREBPs gene expression and their proteolytic activation (Fig. 7) (31-33). Insulin, a potent stimulator of SREBP-1c gene expression, augments triglyceride biosynthesis by inducing transcription of SREBP target genes such as acetyl-CoA carboxylase and fatty acid synthase (34,35). Therefore, one can speculate that insulin may suppress expression of the CYP7A1 gene, a rate-limiting enzyme in bile acid biosynthesis, driven by the activity of LRH-1, probably due to the inhibitory effect of SREBP-1c. Indeed, it has been shown that insulin represses CYP7A1 gene promoter activity in HepG2 cells (36) and that overexpression of SREBP-1c in these cells results in a decline of CYP7A1 gene expression (22,37). On the other hand, when cholesterol is accumulated in the liver, the decrease in active SREBPs impairs the SREBP-mediated suppression of LRH-1 activity, which seems to induce CYP7A1 gene expression. Indeed, we found a slight elevation of CYP7A1 mRNA levels under cholesterol-loaded conditions (Fig. 3B). It is likely that this alternative pathway may play an important role in the regulation of human bile acid biosynthesis, because the human CYP7A1 promoter lacks functional liver X receptorbinding sites (38) critical for the induction of rodent CYP7A1 gene expression in response to hepatic cholesterol accumulation (Fig. 7). Moreover, the SREBP-mediated regulation of SHP gene expression, as shown in the current study, must play a critical role in controlling lipid homeostasis in the liver in light of the fact that SHP affects the transcriptional activities of several nuclear receptors deeply involved in lipid metabolism (10 -15).
The previous study demonstrated that SREBPs directly interact with the ligand-binding domain of HNF-4, thereby inhibiting the recruitment of a coactivator protein, PGC-1␣, and attenuating the transcriptional activity of HNF-4 (39). In the current experiment, SREBPs dramatically diminished the luciferase activity driven by a Gal4DBD-LRH-1 fusion protein, suggesting that the transcriptional activity of LRH-1 is directly suppressed by SREBPs (Fig. 1E). Thus, it can be speculated that the interaction of SREBPs with LRH-1 might interfere with the recruitment of certain LRH-1 coactivators, such as PGC-1␣ and steroid receptor coactivator-1 (25,26) but not the binding to the specific DNA response element. Indeed, we found that SREBP-2 interferes with the PGC-1␣-mediated potentiation of the transcriptional activity of LRH-1 through disrupting the interaction between them and that the N-terminal portion of SREBP-2 is indispensable for this action (Fig. 6). Alternatively, the recent finding, that phosphorylation of the hinge region of LRH-1 close to the ligand-binding domain activates its transcriptional activity, raises the possibility that SREBPs attached to the ligand-binding domain may inhibit the phosphorylation required for LRH-1 transactivation (40). Similarly, it is conceivable that LRH-1 bound to SREBPs might inhibit the recruitment of the coactivators CBP and PGC-1␤, which potentiate SREBP transcriptional activity (24,41). Because it has been reported that SREBP transcriptional activity can be regulated by modifications such as ubiquitylation, sumoylation, acetylation, or phosphorylation (4 -7), we cannot rule out the possibility that disruption of some of these modifications due to the LRH-1 binding to SREBPs may result in the inhibition of the transcriptional activity. These complex molecular mechanisms underlying the mutual inhibition of the transcriptional activities of SREBPs and LRH-1 are now under active investigation. The previous report demonstrated that SREBP-2, but not SREBP-1, down-regulated expression of the sterol 12␣-hydroxylase gene, an LRH-1 target gene, by interacting with LRH-1 (42). Curiously, unlike our current finding that SREBP-1 and -2 equally suppress LRH-1 activity, SREBP-1 adversely activated the sterol 12␣-hydroxylase promoter activity via two SREBPresponsive elements located in the promoter. This discrepancy between the SHP and sterol 12␣-hydroxylase promoters might suggest that 12␣-hydroxylase gene expression is regulated by SREBP-1 more potently than LRH-1 through binding to the SREBP-responsive elements in the promoter sequence, whereas the SHP promoter activity is predominantly controlled by LRH-1. Thus, in the current report we provide the first clear evidence for the detailed molecular interaction between these two proteins, and the mutually inhibitory effects of LRH-1 and the SREBPs on their respective transcriptional activities. A recent study demonstrated that the HMG-CoA reductase promoter, as well as SHP and sterol 12␣-hydroxylase, has functional LRH-1 response elements along with SREBP-binding sites (43). A complex regulation system mediated by these nuclear factors would be predicted to work on these promoters. It seems likely that the cross-talk between SREBPs and nuclear receptors such as HNF-4 and LRH-1 in the liver and small intestine, the most active tissues in lipid metabolism, play pivotal roles for orchestrating their target gene expression in the maintenance of lipid homeostasis. Further studies will be required for elucidating the more complex network among these factors.